Message boards :
AstroPulse :
Refreshed Brook+ app added for testing.
Message board moderation
Author | Message |
---|---|
![]() ![]() Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 ![]() |
There are some changes from last binary so we need to repeat testing. Please, post any issues with cal_ati plan class app here. |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
There are some changes from last binary so we need to repeat testing. Please, post any issues with cal_ati plan class app here. Feeder not running. Can I run that on my 5850?? Anywhere in prefs or app_info or .txt file I need to edit?? |
![]() ![]() Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 ![]() |
There are some changes from last binary so we need to repeat testing. Please, post any issues with cal_ati plan class app here. Run w/o app_info, stock app. At some point CAL support was dropped from Catalyst drivers. So maybe very recent ones are incompatible - we need to check that. |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
oki. I use 12.3 driver. Server still replies "feeder not running" so I'll just have to wait for the switch. |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
Requesting tasks generates a schedular request completed: got 0 tasks. |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
Requesting tasks generates a schedular request completed: got 0 tasks. It doesn't seem to be any new AP's split at the moment for Beta testing, it hasn't been split any new AP's for a long long time here. I may get one or two AP's from time to time, maybe one/day if I'm lucky, and they're always resends _3 or often as high as _7. It's a strange behaviour to start a new test for "refreshed Brook+ app", and not splitting any new AP's for that app. It's like selling a new car, but not providing an engine to the car. :-) |
![]() ![]() Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 ![]() |
I would say more like to sell car working on unavailable fuel :) Will mail Eric about that... |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
I just now received the Brook app and one resent AP (_3) for my HD4850. I'll suspend all other ATI jobs, and will run it immediately. 729 SETI@home Beta Test 2012-12-22 22:09:42 Requesting new tasks for GPU 730 SETI@home Beta Test 2012-12-22 22:09:46 Scheduler request completed: got 1 new tasks 731 SETI@home Beta Test 2012-12-22 22:09:48 Started download of astropulse_6.05_windows_intelx86__cal_ati.exe 732 SETI@home Beta Test 2012-12-22 22:09:48 Started download of brook_cal.dll 733 SETI@home Beta Test 2012-12-22 22:09:52 Finished download of brook_cal.dll 734 SETI@home Beta Test 2012-12-22 22:09:52 Started download of brook.dll 735 SETI@home Beta Test 2012-12-22 22:09:54 Finished download of brook.dll 736 SETI@home Beta Test 2012-12-22 22:09:54 Started download of ap_cmdline_6.04_windows_intelx86__cal_ati.txt 737 SETI@home Beta Test 2012-12-22 22:09:55 Finished download of ap_cmdline_6.04_windows_intelx86__cal_ati.txt 738 SETI@home Beta Test 2012-12-22 22:09:55 Started download of astropulse_6.05_cal_ati_AUTHORS 739 SETI@home Beta Test 2012-12-22 22:09:56 Finished download of astropulse_6.05_cal_ati_AUTHORS 740 SETI@home Beta Test 2012-12-22 22:09:56 Started download of astropulse_6.05_cal_ati_COPYING 741 SETI@home Beta Test 2012-12-22 22:09:57 Finished download of astropulse_6.05_cal_ati_COPYING 742 SETI@home Beta Test 2012-12-22 22:09:57 Started download of astropulse_6.05_cal_ati_COPYRIGHT 743 SETI@home Beta Test 2012-12-22 22:09:59 Finished download of astropulse_6.05_cal_ati_COPYRIGHT 744 SETI@home Beta Test 2012-12-22 22:09:59 Started download of ap_22my12ad_B4_P0_00147_20121103_24020.wu 745 SETI@home Beta Test 2012-12-22 22:10:16 Finished download of ap_22my12ad_B4_P0_00147_20121103_24020.wu 746 SETI@home Beta Test 2012-12-22 22:10:32 Finished download of astropulse_6.05_windows_intelx86__cal_ati.exe |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
Hmm, why did it send me ap_cmdline_6.04_windows_intelx86__cal_ati.txt, when the version should be 6.05 according to the other files, as well as the application named "astropulse_v6 version 605" (6.05 AstroPulse v6 (cal_ati) Edit, added: GPU load is ridiculously low (I haven't run any brook apps on this GPU before, so maybe that's normal). Mostly nothing, sometime 1%, and only when changing percentage done, it jumps up to 50-60%, during the transition of percentage done (a few seconds only). This WU is 0% blanked <fraction_blanked>0.000000</fraction_blanked> |
![]() ![]() Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 ![]() |
Hmm, why did it send me ap_cmdline_6.04_windows_intelx86__cal_ati.txt, when the version should be 6.05 according to the other files, as well as the application named "astropulse_v6 version 605" (6.05 AstroPulse v6 (cal_ati) 1) Another question to Eric :) I think just typo. 2) Such load is normal. That 1% actually not from app but from rest of your system. App GPU usage between FFA is absolute zero. But for CAL-only GPUs it's better than nothing. Should not be used in wild for OpenCL-capable ones, of course. EDIT: Total GPU load can be increased running few tasks at once though. BTW such approach is very similar how GPU itself achieve high performance in spite of big memory latency. It launches many waves so while one wave awaiting for memory (in our case - one task doing CPU-only work) another uses GPU. |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
Hmm, why did it send me ap_cmdline_6.04_windows_intelx86__cal_ati.txt, when the version should be 6.05 according to the other files, as well as the application named "astropulse_v6 version 605" (6.05 AstroPulse v6 (cal_ati) I see, but without an app_info (running stock) it would not be possible to run more than one task at a time I guess. Well, since my HD4850 is OpenCL capable, I have only been running OpenCL apps on it so far. I guess it's normal here on Beta though, that it also got this Brook app, and that it would not happen on main, for an OpenCL capable GPU? |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
I see, but without an app_info (running stock) it would not be possible to run more than one task at a time I guess. There's a new app_config.xml option in Boinc 7.0.42 where you can do just that: BOINC 7.0.40-42 and new app_config.xml Claggy |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
Hmm, seems like I got only 3 "normal" astropulse tasks... http://setiweb.ssl.berkeley.edu/beta/results.php?hostid=55958 |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
I see, but without an app_info (running stock) it would not be possible to run more than one task at a time I guess. But you know that I refuse to upgrade to any Boinc 7 version, from a perfectly working 6.10.58. So I guess I will not run more than one Brook AP... The day I upgrade to version 7, is probably the real date when the world ends :-) |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
Are the cal tasks considered CPU tasks or ATI tasks at the prefs for seti beta?? I have noticed other projects in conflict with BM about wheather the task is considered a CPU or a GPU task.... That is noticeble when limiting 1 core to feed GPU and when the Einstein/POEM task is started another CPU core is lost.... But that is a BM problem... And I don't do BM Alpha tests anymore... |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
Are the cal tasks considered CPU tasks or ATI tasks at the prefs for seti beta?? Must be considered as GPU tasks, since when my computer asked for GPU tasks, it got one Brook WU, and the app files. 729 SETI@home Beta Test 2012-12-22 22:09:42 Requesting new tasks for GPU 730 SETI@home Beta Test 2012-12-22 22:09:46 Scheduler request completed: got 1 new tasks 731 SETI@home Beta Test 2012-12-22 22:09:48 Started download of astropulse_6.05_windows_intelx86__cal_ati.exe (snipped the other lines in this request) |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
There should be an option for cal tasks in "seti beta prefs"/"seti prefs". And they should count as ATI GPU tasks. |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
Are the cal tasks considered CPU tasks or ATI tasks at the prefs for seti beta?? I hope so. I'll see tomorrow if I get any cal tasks. I do have cpu unmarked in seti beta prefs...Yours is marked though. |
![]() Send message Joined: 10 Mar 12 Posts: 1700 Credit: 13,216,373 RAC: 0 ![]() |
Are the cal tasks considered CPU tasks or ATI tasks at the prefs for seti beta?? Oh yeah, I have CPU as well as GPU marked. Unmarked V7 and SETI@home v7 large workunits, as well as "If no work for selected applications is available, accept work from other applications?" though. This setting is just to maximise the chance of getting AP's. Normally I do accept every kind of WU I can get, but since there are so few AP's now, unmarking the other versions seems the best option for now. |
![]() ![]() Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 ![]() |
I guess it's normal here on Beta though, that it also got this Brook app, and that it would not happen on main, for an OpenCL capable GPU? Depends on BOINC server config... I hope that config will be "never send brook app if any of OpenCL ATi plans supported". We already know that BOINC can't select fastest app (checked with CUDA apps). So we can't rely on this feature. Don't know if some other way exists to implement such config. If not I would hold release on main cause if HD7xxx will start to recive brook+ tasks it will be big overall slowdown instead of speedup. |
©2021 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.